| Message ID | 1560325326-1598-1-git-send-email-stanley.chu@mediatek.com (mailing list archive) |
|---|---|
| State | Mainlined, archived |
| Commit | 24e2e7a19f7e4b83d0d5189040d997bce3596473 |
| Headers | show |
| Series | [v1] scsi: ufs: Avoid runtime suspend possibly being blocked forever | expand |
Hi, > > UFS runtime suspend can be triggered after pm_runtime_enable() > is invoked in ufshcd_pltfrm_init(). However if the first runtime > suspend is triggered before binding ufs_hba structure to ufs > device structure via platform_set_drvdata(), then UFS runtime > suspend will be no longer triggered in the future because its > dev->power.runtime_error was set in the first triggering and does > not have any chance to be cleared. > > To be more clear, dev->power.runtime_error is set if hba is NULL > in ufshcd_runtime_suspend() which returns -EINVAL to rpm_callback() > where dev->power.runtime_error is set as -EINVAL. In this case, any > future rpm_suspend() for UFS device fails because > rpm_check_suspend_allowed() fails due to non-zero > dev->power.runtime_error. > > To resolve this issue, make sure the first UFS runtime suspend > get valid "hba" in ufshcd_runtime_suspend(): Enable UFS runtime PM > only after hba is successfully bound to UFS device structure. > > Fixes: e3ce73d (scsi: ufs: fix bugs related to null pointer access and array size) This code was inserted before platform_set_drvdata in 6269473 [SCSI] ufs: Add runtime PM support for UFS host controller driver. Why do you point to e3ce73d? Thanks, Avri > Cc: stable@vger.kernel.org > Signed-off-by: Stanley Chu <stanley.chu@mediatek.com> > --- > drivers/scsi/ufs/ufshcd-pltfrm.c | 11 ++++------- > 1 file changed, 4 insertions(+), 7 deletions(-) > > diff --git a/drivers/scsi/ufs/ufshcd-pltfrm.c b/drivers/scsi/ufs/ufshcd-pltfrm.c > index 8a74ec30c3d2..d7d521b394c3 100644 > --- a/drivers/scsi/ufs/ufshcd-pltfrm.c > +++ b/drivers/scsi/ufs/ufshcd-pltfrm.c > @@ -430,24 +430,21 @@ int ufshcd_pltfrm_init(struct platform_device > *pdev, > goto dealloc_host; > } > > - pm_runtime_set_active(&pdev->dev); > - pm_runtime_enable(&pdev->dev); > - > ufshcd_init_lanes_per_dir(hba); > > err = ufshcd_init(hba, mmio_base, irq); > if (err) { > dev_err(dev, "Initialization failed\n"); > - goto out_disable_rpm; > + goto dealloc_host; > } > > platform_set_drvdata(pdev, hba); > > + pm_runtime_set_active(&pdev->dev); > + pm_runtime_enable(&pdev->dev); > + > return 0; > > -out_disable_rpm: > - pm_runtime_disable(&pdev->dev); > - pm_runtime_set_suspended(&pdev->dev); > dealloc_host: > ufshcd_dealloc_host(hba); > out: > -- > 2.18.0
Hi Avri, On Wed, 2019-06-12 at 11:10 +0000, Avri Altman wrote: > Hi, > > > > > UFS runtime suspend can be triggered after pm_runtime_enable() > > is invoked in ufshcd_pltfrm_init(). However if the first runtime > > suspend is triggered before binding ufs_hba structure to ufs > > device structure via platform_set_drvdata(), then UFS runtime > > suspend will be no longer triggered in the future because its > > dev->power.runtime_error was set in the first triggering and does > > not have any chance to be cleared. > > > > To be more clear, dev->power.runtime_error is set if hba is NULL > > in ufshcd_runtime_suspend() which returns -EINVAL to rpm_callback() > > where dev->power.runtime_error is set as -EINVAL. In this case, any > > future rpm_suspend() for UFS device fails because > > rpm_check_suspend_allowed() fails due to non-zero > > dev->power.runtime_error. > > > > To resolve this issue, make sure the first UFS runtime suspend > > get valid "hba" in ufshcd_runtime_suspend(): Enable UFS runtime PM > > only after hba is successfully bound to UFS device structure. > > > > Fixes: e3ce73d (scsi: ufs: fix bugs related to null pointer access and array size) > This code was inserted before platform_set_drvdata in > 6269473 [SCSI] ufs: Add runtime PM support for UFS host controller driver. > Why do you point to e3ce73d? e3ce73d (scsi: ufs: fix bugs related to null pointer access and array size) changed the returned value from 0 to -EINVAL in case of NULL "hba" in ufshcd_runtime_suspend(). But you are right, above patch may do the right thing, and the real root cause is the incorrect timing of pm_runtime_enable(). I will fix commit message in next version. > > Thanks, > Avri Thanks. Stanley
diff --git a/drivers/scsi/ufs/ufshcd-pltfrm.c b/drivers/scsi/ufs/ufshcd-pltfrm.c index 8a74ec30c3d2..d7d521b394c3 100644 --- a/drivers/scsi/ufs/ufshcd-pltfrm.c +++ b/drivers/scsi/ufs/ufshcd-pltfrm.c @@ -430,24 +430,21 @@ int ufshcd_pltfrm_init(struct platform_device *pdev, goto dealloc_host; } - pm_runtime_set_active(&pdev->dev); - pm_runtime_enable(&pdev->dev); - ufshcd_init_lanes_per_dir(hba); err = ufshcd_init(hba, mmio_base, irq); if (err) { dev_err(dev, "Initialization failed\n"); - goto out_disable_rpm; + goto dealloc_host; } platform_set_drvdata(pdev, hba); + pm_runtime_set_active(&pdev->dev); + pm_runtime_enable(&pdev->dev); + return 0; -out_disable_rpm: - pm_runtime_disable(&pdev->dev); - pm_runtime_set_suspended(&pdev->dev); dealloc_host: ufshcd_dealloc_host(hba); out:
UFS runtime suspend can be triggered after pm_runtime_enable() is invoked in ufshcd_pltfrm_init(). However if the first runtime suspend is triggered before binding ufs_hba structure to ufs device structure via platform_set_drvdata(), then UFS runtime suspend will be no longer triggered in the future because its dev->power.runtime_error was set in the first triggering and does not have any chance to be cleared. To be more clear, dev->power.runtime_error is set if hba is NULL in ufshcd_runtime_suspend() which returns -EINVAL to rpm_callback() where dev->power.runtime_error is set as -EINVAL. In this case, any future rpm_suspend() for UFS device fails because rpm_check_suspend_allowed() fails due to non-zero dev->power.runtime_error. To resolve this issue, make sure the first UFS runtime suspend get valid "hba" in ufshcd_runtime_suspend(): Enable UFS runtime PM only after hba is successfully bound to UFS device structure. Fixes: e3ce73d (scsi: ufs: fix bugs related to null pointer access and array size) Cc: stable@vger.kernel.org Signed-off-by: Stanley Chu <stanley.chu@mediatek.com> --- drivers/scsi/ufs/ufshcd-pltfrm.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-)